How to use ReXML "in the wild"?


#1

I’d very much like to use ReXML’s XPATH features to extract info from
Google’s financial info pages, but find that Rexml chokes on the
Javascript, here’s the result of trying to read in a page with this
bit of code:

require “rexml/document”
require ‘net/http’
Net::HTTP.start(‘finance.google.com’) do |http|
response = http.get(’/finance?fstype=ii&q=NYSE:WAT’)
rdoc = REXML::Document.new(response.body)
end

==========
Output:

/usr/local/lib/ruby/1.8/rexml/parsers/treeparser.rb:92:in parse': #<RuntimeError: Illegal character '&' in raw string " (REXML::ParseException) (function(){ var d=navigator.userAgent.toLowerCase().indexOf("msie")!=-1;function e(){var b=document.styleSheets;for(var a=b.length-1;a>=0;--a){var c=b[a].href;if(c)if(c.indexOf("styles/finance_")!=-1|| c.indexOf("styles_")!=-1)return b[a]}return null}function f(){var b=e();if(b){var a=b.rules;return a.length>0&&a[a.length-1].selectorText==".lastFinanceRule"}return false} function g(){if(document.scripts)for(var b=0;b"> /usr/local/lib/ruby/1.8/rexml/text.rb:91:ininitialize’
.
.
.

Is there a good way to get around this problem? If, not, I guess it’s
back to regular expressions…

Thanks,
Ken


#2

Hi Kenneth,

I’d very much like to use ReXML’s XPATH features to extract info from
Google’s financial info pages, but find that Rexml chokes on the
Javascript, here’s the result of trying to read in a page with this
bit of code:

Don’t try that :wink: REXML in the wild == epic FAIL. At this level, you
might
want to try Hpricot or Nokogiri. At a bit higher level, scRUBYt!
You can read about web scraping in Ruby here (my most succesfull article
ever, was even mentioned in Learning R. from O’Reilly):

http://www.rubyrailways.com/data-extraction-for-web-20-screen-scraping-in-rubyrails/

Is there a good way to get around this problem? If, not, I guess it’s
back to regular expressions…

Web scraping with regular expressions is almost never a good idea.

Try scRUBYt!:

require ‘rubygems’
require ‘scrubyt’

data = Scrubyt::Extractor.define do
fetch ‘http://finance.google.com/finance?fstype=ii&q=NYSE:WAT

body ‘/html/body’ do
revenue ‘/div[4]/div[2]/table/tr[2]’ do
ending_9_27 ‘/td[2]’
ending_6_28 ‘/td[3]’
end

gross_profit '/div[4]/div[2]/table/tr[2]' do
  ending_9_27 '/td[2]'
end

end
end

puts data.to_xml

output:

386.31 398.77 386.31

HTH,
Peter


http://scrubyt.org
http://www.rubyrailways.com


#3

Kenneth McDonald wrote:

I’d very much like to use ReXML’s XPATH features to extract info from
Google’s financial info pages, but find that Rexml chokes on the
Javascript, here’s the result of trying to read in a page with this
bit of code:

I have studied REXML for many years, and I still can’t figure out how to
get it
to recognize an — or similar advanced entity.

Like the other responder said, give up while you still can. libxml-ruby
is also
stable enough to give a shot - oh yeah, except it crashes on non-tiny
inputs.

Aaaand…

/usr/local/lib/ruby/1.8/rexml/parsers/treeparser.rb:92:in `parse’:
#<RuntimeError: Illegal character ‘&’ in raw string

That’s because REXML and your web browser disagree on the definition of
well-formed. Your browser accepts a naked & inside a JavaScript tag, but
REXML
does not. REXML is technically correct, and your browser would have
accepted
&& here, but…

a.length>0&&a[a.length-1].selectorText==".lastFinanceRule"}return false}

…browsers cannot correctly interpolate & appearing inside JavaScript
literal
strings, because some lowlife coder using Notepad might have actually
wanted
“&” when they wrote “&” - such as with document.write().

So, because REXML cannot accept normal HTML, due to hits and misses of
standards
compliance on all sides - you are better off with a dedicated parser!